To help Konrad and keyboardP resolve their differences, I ran a benchmark test, using their code. It turns out that keyboardP's code is 10x faster than Konrad's code
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Text.RegularExpressions;
namespace ConsoleApplication1
{
class Program
{
static void Main(string[] args)
{
string input = "asdf234!@#*advfk234098awfdasdfq9823fna943";
DateTime start = DateTime.Now;
for (int i = 0; i < 100000; i++)
{
RemoveNonUnicodeLetters(input);
}
Console.WriteLine(DateTime.Now.Subtract(start).TotalSeconds);
start = DateTime.Now;
for (int i = 0; i < 100000; i++)
{
RemoveNonUnicodeLetters2(input);
}
Console.WriteLine(DateTime.Now.Subtract(start).TotalSeconds);
}
public static string RemoveNonUnicodeLetters(string input)
{
StringBuilder sb = new StringBuilder();
foreach (char c in input)
{
if (Char.IsLetter(c))
sb.Append(c);
}
return sb.ToString();
}
public static string RemoveNonUnicodeLetters2(string input)
{
var result = Regex.Replace(input, "\\P{L}", "");
return result;
}
}
}
I got
0.12
1.2
as output
UPDATE:
To see if it is the Regex compilation that is slowing down the Regex method, I put the regex in a static variable that is only constructed once.
static Regex rex = new Regex("\\P{L}");
public static string RemoveNonUnicodeLetters2(string input)
{
var result = rex.Replace(input,m => "");
return result;
}
But this had no effect on the runtime.